Human Capital and
Regional Development
Nicola Gennaioli
Rafael La Porta
Florencio Lopez-de-Silanes
Andrei Shleifer
May 2011
Motivation (1)

Competing views on the ultimate determinants of economic development:
1.
2.
3.
4.
5.
Geography – Bloom and Sachs (1998);
Institutions – King & Levine (1993), De Long & Shleifer (1993), Acemoglu et al.
(2001);
Human Capital – Lucas (1988), Barro (1991), Mankiw, Romer & Weil (1992);
Ethnic heterogeneity -- Easterly & Levine (1997), Alesina et al. (2003); and
Culture – Knack & Keefer (1997).

These variables are correlated with each other and, in particular, with human capital.
 Difficult to disentangle the ultimate determinants of economic development.
 Instrumental variable techniques are not helpful.

We collect regional (i.e. sub-national) data to shed light on these debates.
 We run GDP level regressions (Hall & Jones, 1999) but also regressions using
establishment-level data in these regions.
2
Motivation (2)

To organize the discussion, we present a new model of regional development.
 The human capital of workers enters as an input into the neoclassical production
function, but the entrepreneur’s human capital independently influences firm-wide
productivity (Lucas 1978);
 Human capital may have externalities (Lucas 1988, 2008); and
 Workers and entrepreneurs may move across regions (e.g., Glaeser and Gottlieb 2009).

We find:
 Strong evidence that geography matters.
 No evidence that institutions, ethnic, culture matter but these may be data problems.
 Overwhelming evidence that human capital fosters development through
entrepreneurial education and externalities.
3
Outline

Model

Empirical Predictions

Data

Results

Conclusion
4
The Lucas-Lucas Model (1)

Each country consists of a measure 1 of regions.
 Regions are denoted by i.
 p: share of “productive regions” (productivity AG).
 1–p: share of “unproductive” regions (productivity AB < AG).

A measure 2 of agents is uniformly distributed across regions.
 Agents are denoted by j.
 They enjoy consumption (c) and housing (a) according to the utility function:
[1]

1 j  j
u ( c, a )  c
a
Half the agents are “rentiers,” the remaining half are “labourers’’.
 Each rentier owns 1 unit of housing, T units of land and K units of physical capital.
Rentiers have no human capital and do not consume housing (θj=0).
 Each labourer has h units of human capital.
 h has a Pareto distribution with support [h,+∞) and mean E(h) = μh/(μ–1).
5
The Lucas-Lucas Model (2)

A labourer can become either an entrepreneur or a worker.
 By operating in region i, an entrepreneur with human capital h who hires physical
capital Ki,h , land Ti,h , and workers with total human capital Hi,h produces an amount
of the consumption good equal to:
[2] yi,h  Ai h1   Hi,h Ki,hTi,h



where α+β+δ<1 and output increases at a diminishing rate with h (Lucas 1978).
Entrepreneurs earn profits πi(h).
Workers earn wage income wi·h as a worker, where wi is the wage rate

Rentiers rent land and physical capital to firms, and housing to labour.
 In region i, a representative rentier earns λiT and ηi by renting land and housing.
 A rentier renting physical capital in region i earns ρiK.

Physical capital is fully mobile across regions; land and housing endowments are fixed;
Labourers are partially mobile: lose φwi units of income by moving to region i (φ <h).
The Lucas-Lucas Model (3)

There are two time periods (0 and 1).
 Time 0.
1.
Each labourer selects the location and occupation that maximize her income.
2.
The housing market clears.
 Time 1.
1.
Entrepreneurs hire land, human, and physical capital.
2.
Production is carried out and distributed (i.e. wi, πi, λi, ηi, and ρi are paid).
3.
Consumption takes place.
7
Equilibrium (1)

There is a unique rental rate ρ (physical capital is fully mobile) .

Rental rates λi and ηi vary across regions depending on productivity and population.

The sorting of labourers into workers and entrepreneurs is determined by wi and πi(hj).
 An entrepreneur with human capital h operating in region i solves:
[3]

H i ,h ,Ti ,h , K i ,h
Firm j employs a share of entrepreneurial capital hj/HiE and thus hires the others
factors according to:
[4]

max Ai h1    H i,h K i,hTi ,h  wi H i ,h  K i ,h  iTi ,h
Hi, j 
hj
H iE
 H iW , Ki , j 
hj
H iE
 Ki , Ti , j 
hj
H iE
T.
Which implies that aggregate regional output is given by
[5]
     H  K  T 
Yi  Ai H iE
1  
W
i
i
8
Equilibrium (2)

Based on the equation for aggregate regional output, wages, profits, and capital rental
rates are given by:
[6]



K / H  T / H  ,
wi 
Yi
   Ai H iE / H iW
W
H i
i 
Yi
 (1       )  Ai H iW / H iE
E
H i

Yi
   Ai H iE / K i
K i
1   



1   
W
i
i
H
W
i
W
i
 K / H  T / H  ,
/ Ki
i
E
i
E
i
 T / K  .
i
In equilibrium, πi = wi so that labourers are indifferent between occupations:
[7 ]

1       
 

H iE  
  H i , H iW  
  Hi
 1    
1     
where Hi=HiE+HiW is total human capital in region i.
9
Lucas-only spatial equilibrium (1)

We consider a symmetric spatial equilibrium:
 All productive regions share the same factor allocation (HG,KG), the same wage wG and
rental rates λG and ηG, and
 All unproductive regions share the same factor allocation (HB,KB), wage rate wB, and
rental rates λB and ηB.

The housing rental rate is equal to the share θ of labour income in the region. As a
consequence, the utility of a labourer from staying in his birth region i is equal to:


wi1
h
i
  H i
Utility rises with wi and falls with regional human capital Hi due to higher rental rates.
Formally, a labourer with human capital h migrates if:
[11]
uw,i (c, a) 
wi h


wG1 ( h   ) / H G  w1B h / H B

where φ captures migration costs. This identifies a human capital threshold hm such
10
that agent j migrates if and only if hj ≥ hm.
Lucas-only spatial equilibrium (2)

To simplify national aggregation, we study the case where mobility costs are negligible. In
this case:
1. National output is given by:
[19]
^
     H  K  T 
Y  A HE
1  
W


where A is a function A(  ,  , , AG , AB , p) of exogenous parameters.
2.
Regional output is described by [5]:
[5]
3.
     H  K  T 
Yi  Ai H iE
1  
W
i
i
Firm-level output is given by [2].
[2]
yi,h  Ai h1   Hi,h Ki,hTi,h
11
Lucas-Lucas spatial equilibrium (1)

We assume externalities such that regional total factor productivity equals to:
[20]









~
Ai  Ai Ei (h) Li ,
  0,  1
Ei(h) is the average level of human capital in region i.
Li is the measure of labourers in that region.
ψ captures the importance of the quality of human capital.
 ψ = 1 only the total quantity of human capital Hi = Ei(h)Li matters for externalities;
 ↑ψ: the quality of human capital becomes relatively more important than quantity.
γ captures the overall importance of externalities.
Under perfect mobility, productive regions employ human capital in proportion to their
1
relative productivity:
HG
AG (1 ) (1 )
~*
[17] H G 

1
E ( h)


 (1 )  (1 )

EA


12
~
~*
With mobility costs, H G settles between 1 (no mobility) and H G (perfect mobility).
Lucas-Lucas spatial equilibrium (2)

The sorting of human capital across regions depends on the difference β – γ between the
diseconomies of scale due to the presence of a fixed factor and the positive human capital
spillovers.
 The smaller is β– γ, the stronger are the spillovers and the greater is the migration of
human (and physical) capital into the more productive regions.

A stable equilibrium requires that:
[22]
(β – γ)(1 – θ) + θ(1 – δ) > 0
 the share θ of income spent by labourers on housing must be sufficiently large if γ > β.
 Intuition: Migration increases productivity by creating spillovers but also increases the
housing and land rental rates so that after some point migration stops.
13
Lucas-Lucas spatial equilibrium (3)

In a stable equilibrium:
 National output is equal to:
[23]
^
     H  K  T 
Y  AH H E
1  
W


where A is a function A(  ,  , , AG , AB , p) of exogenous parameters.

Firm level output is given by:
[24]

yi , j  Ai Ei (h) Li h1j   Hi, j Ki, jTi,j
Regional output becomes:
[25]
Yi  Ai Ei (h) Li ( H iE )1   ( H iW ) Ki T 
14
Empirical Predictions
15
Measurement of Human Capital

To implement the model, we need to measure human capital based on schooling (“S”).
 We follow the Mincerian approach in which for an individual j the link between
human capital and schooling is:
[ 26]

h j  exp j S j 
where Sj ≥ 0 and μj ≥ 0 are two random variables.
16
Regional Income Differences

Regional output is given by:
[28]


ln(Yi/Li) = C + (1/(1 – δ))lnAi + (1+ γψ – β/(1 – δ))lnEi(h) + (γ – β/(1 – δ))lnLi
where C is a constant absorbed by the country fixed effect.
Using the Mincerian specification for human capital, regional output is:
1


lnAi  (1   
) i Si  ( –
)lnLi
1- 
1- 
1- 
 Coefficient on average regional schooling S i  product of the technological parameter
(1+ γψ – β/(1 – δ)) and the nation-wide average Mincerian return  .
 Fixed land supply progressively reduces the productivity of human capital.
[29]
ln(Yi /Li )  C 
 Population Li on the right hand side captures the productive return to increasing regional
population (or workforce), which is equal to (γ – β/(1 – δ)).
17
 When population rises, productivity rises because the total quantity of human capital
rises but it falls due to fixed land supply.
Firm-level productivity (1)

Equation [24] implies that firm-level output per worker yi,j/li,j can be expressed as:
[30]
ln(yi, j/li, j )  lnAi  (1 –  -  -  )ln[ Ei,j(h E )(li,jE / li,j )]  α ln[(hW )(li,jW /li,j )] 
δ ln ki,j  β ln ti,j  γ ln Li  γ ln Ei (h)


where xi,j = Xi,j/li,j denote per-worker values.
Defining  E ,i and  W ,i as the average Mincerian return of entrepreneurs and workers in
region i, the empirical counterpart of Equation [30] then becomes:
[32]
ln(yi, j/li, j )  lnAi  (1 –  -  -  )  E ,iSE,ij  W ,iSW,ij  (1 –  -  -  )ln(lijE /lij ) 
ln(lWij /lij )  lnkij   ln tij  lnLi  i Si

Key property of our analysis: Separate the remuneration α of “low human capital” labour,
and the remuneration (1–α–β–δ) of “high human capital”.
 Sorting implies that more talented people become entrepreneurs and less talented
people become workers. Thus,   
18
E ,i
W ,i
Firm-level productivity (2)

Coefficient on SE,ij  entrepreneurs’ rents (1–α–β–δ) times the average nation-wide
Mincerian return of entrepreneurs.

Coefficient on SW,ij  labour share α times the average nation-wide Mincerian return of
workers.

Coefficient on regional schooling  externality parameter γψ times the population-wide
average Mincerian return.
19
Data
20
The Variables

We found regional (i.e. sub-national) data on either income or education for 110 countries.
 For those 110 countries, in addition to income and education, we collected data on:
1. Geography and endowments.
1. Temperature,
2. Inverse distance to coast, and
3. Oil.
2. Institutions
1.
2.
3.
4.
5.
6.
7.
8.
Informal payments,
Days to pay taxes,
Days without electricity,
Security costs,
Access to land,
Access to finance,
Government predictability, and
Doing Business rank
3.
Infrastructure
1. Power line density, and
2. Time to travel to the closest city
of 50,000 inhabitants.
4. Culture
1.
2.
3.
4.
Trust,
Civic values,
Number of ethnic groups, and
Probability of same language.
5. Population
21
Regions

Countries have administrative divisions. In turn, administrative divisions may have
different levels (e.g., country  states or provinces  counties or municipalities).

For each variable, we collect data at the highest administrative division available or, when
such data does not exist, at the statistical division (e.g. Eurostat NUTS) that is closest to it.

We aggregate data for each country to a region from the most disaggregated level of
reporting available.
 Ex: Brazil has GDP and education data for 27 first-level administrative regions and
Enterprise Survey data for 432 municipalities. We aggregate the Enterprise Survey
data by averaging all municipalities within the same first-level administrative division.

The final dataset has 1,569 regions.
 79 countries at the first-level administrative division;
 31 countries at a more aggregated level than the first-level administrative division.
 Often because education is unavailable at the first-level administrative level.
 Ex: Ireland publishes GDP data for 8 regions but education data for 2 regions.
22
Coverage

The countries in our sample account for 97% of the world’s GDP in 2005.

Coverage is related to:
1.
Surface area, presumably because very small countries do not report regional data.
We only have data for 14% of the smallest 50 countries
2.
Absolute level of GDP but not with GDP per capita. We have data for 18% the first
23
50 countries in terms of GDP in 2005, but 52% on the basis of GDP per capita.
Descriptive Statistics
Table 1 – Panel A
Medians for:
Number of Number of Regions per Regions
Countries
country
Ln(GDP per capita)
Years of Education
Temperature
Inverse Distance to Coast
Ln(Oil)
Informal Payments
ln(Tax Days)
Ln(Days without electricity)
Security costs
Access to land
Access to finance
Government Predictability
Doing Business Percentile Rank
Ln(Power line density)
Ln(Travel time)
Trust in others
Civic Values
Ln(Number of ethnic groups)
Probability of same language
1,537
1,489
1,568
1,569
1,569
361
270
222
373
519
536
386
180
1,569
1,569
745
683
1,568
1,545
107
106
110
110
110
76
58
75
79
81
82
75
19
110
110
69
75
110
109
11
12
12
12
12
4
5
2
4
5
5
4
6
12
12
9
8
12
12
Mean
Minimum
Maximum
8.69
6.58
16.84
0.90
0.00
1.02
1.29
3.03
0.91
0.15
0.28
0.46
0.40
1.34
5.28
0.23
2.23
0.98
0.67
8.07
5.34
10.23
0.80
0.00
0.40
1.06
2.73
0.39
0.04
0.14
0.34
0.21
0.00
4.21
0.12
1.71
0.00
0.28
9.54
8.70
21.13
0.99
0.00
1.60
1.51
3.37
1.34
0.27
0.47
0.61
0.49
2.53
6.00
0.38
3.12
1.79
0.79
Within‐country Within‐country Coefficient of Variation
Range
std deviation for Variable in Levels
1.03
2.34
4.47
0.13
0.00
0.94
0.36
0.54
0.72
0.21
0.29
0.24
0.22
1.87
1.82
0.22
1.08
1.39
0.26
0.30
0.73
1.45
0.05
0.00
0.45
0.19
0.36
0.34
0.09
0.12
0.10
0.11
0.61
0.54
0.07
0.48
0.50
0.09
0.33
0.92
0.09
0.05
0.00
0.59
0.18
0.32
0.42
0.40
0.24
0.20
0.31
0.61
0.46
0.35
0.19
0.46
0.21
24
Standard deviation of Ln Regional GDP per capita)
-.2
0
.2
.4
.6
Within-country standard deviation of GDP
per capita and development -- Figure 4
THA
KEN
ZAR
IDN
ARG
PAN
MOZ
RUS
PER IRN
NAM
CHN
BRA
ARE
COL
KGZ
IND
GHA
MNG
NGA
LVA MEX
CHL
SVK
NER BEN UZB
MKDTUR
HNDBOL
ALBECU
MYS
ZAF
UGA
UKR
GTM SRB ROM
LKA
EGY KAZ HUN
VNM
ZMB
LSO
TZA
LTU CZEBEL
BLZ VEN
EST
IRL
DOM
PRY BIH
LBN
MDA
DEU
CAN
CHE
NIC
MDG KHM
HRV GABITA
CMR
USA
LAO PHL SWZ URY
SLV
NOR
GRC
PRT
AUT
FIN
BGRPOL KOR
DNK
NZL
SVN
MAR
GEO
GBR
NPL
ESP
NLD
JPN
AUS
SWE
ARM
MWI BFASEN
JOR
FRA
ISR
SYR
PAK
AZE
ZWE
-6
-4
-2
Ln GDP per capita
0
2
coef = -.02489678, (robust) se = .01172037, t = -2.12
25
1.5
Within-country standard deviation of years of
education and development – Figure 5
PAN
CMR
NGA
GHA
ZAR
ZWE
-.5
Std Dev Years of Education
0
.5
1
KEN
-6
-4
THA
PER
LBN
NAM
IDN
IND PRY
LAOHND
COL
BLZ
GAB
BEN NIC
MDA GTM
MYS
MDG
CHN
SWZ
ISR
DOM
MEX
PHL EGY
ECU
ARG
BRA
KHM
SRB
SYR
NER
ZAF
VNM BOL
LSOSEN
TZA
BFA
SVK
MNG
MKD
CHL
LVA
ARE
VEN
HRV
TUR
JOR URY
MAR
ROM
BGR
UGA
SLV
NZL
MOZ
DNK
DEU
AUS
GEO
UKR
ESP
MWI
ARM
LTU
HUNGRC
JPN
EST CZE
LKA AZE RUS
BEL
PAK
PRT
FRA
SVN
KAZ
IRL
KGZ
NOR
BIH
CAN
GBR
ITA
AUT
UZB
NLD
SWE
USA
NPL
FIN
POL
CHE
ZMB
-2
Ln(GDP per capita)
0
2
coef = -.12647737, (robust) se = .02655853, t = -4.76
26
Descriptive Statistics
Table 1– Panel B
Medians for:
Number of Number of Regions per Regions
Countries
country
Ln(Establishments / Population)
Ln(Employees / Establishments)
Ln(Employees / Population)
Ln(Employees Big Firms / Employees)
Ln(Sales / Employee)
Ln(Wages / Employee)
Ln(Employees)
Ln(Expenditure on energy / Employee)
Ln(Property, plant and equipment / Employee)
Years of Education of Workers
Years of Education of Managers
984
1,068
1,056
540
550
516
550
326
205
507
195
65
69
69
31
82
77
82
66
41
74
38
12
12
12
13
5
5
5
4
4
5
4
Mean
‐4.89
2.07
‐2.66
‐1.45
10.21
8.28
3.25
6.10
8.72
9.97
14.90
Minimum Maximum
‐5.45
1.69
‐3.38
‐2.17
9.79
8.00
2.72
5.52
8.37
8.66
14.24
‐4.06
2.39
‐1.80
‐0.78
10.59
8.66
3.71
6.36
9.37
10.80
15.36
Within‐
Within‐country Coefficient of Variation country Range std deviation
for Variable in Levels
1.17
0.80
1.58
1.13
0.79
0.62
0.82
0.60
0.99
2.25
1.34
0.37
0.20
0.43
0.33
0.35
0.25
0.35
0.30
0.47
0.93
0.62
0.37
0.19
0.41
0.27
1.22
1.79
1.46
1.22
1.26
3.06
0.89
27
Results
28
Univariate Fixed Effect Regressions
Table 3
Independent Variables:
Years of Education
Temperature
Inverse Distance to Coast
Ln(Oil)
Informal Payments
ln(Tax Days)
Ln(Days without electricity)
Security costs
Access to land
Access to finance
Government Predictability
Doing Business Percentile Rank
Ln(Power line density)
Ln(Travel time)
Trust in others
Ln(Number of ethnic groups)
Probability of same language
Observations
Countries
R2 Within
R2 Between
1,470
1,536
1,537
1,537
350
263
219
362
507
524
380
176
1,537
1,537
739
1,536
1,518
104
107
107
107
74
56
73
77
79
80
73
18
107
107
68
107
106
38%
1%
4%
2%
0%
0%
2%
0%
0%
1%
1%
2%
5%
7%
0%
5%
1%
58%
27%
13%
4%
21%
20%
6%
7%
15%
8%
0%
13%
36%
15%
18%
17%
26%
29
National GDP Per Capita and Geography
Table 4 – Panel A
(1)
(2)
a
(3)
Temperature
‐0.0914
(0.0100)
‐0.0189
(0.0106)
c
‐0.0190
(0.0106)
Inverse Distance to Coast
4.4768
(0.5266)
a
2.9647
(0.5736)
a
2.9499
(0.5782)
Ln(Oil)
1.2192
(0.1985)
a
0.9503
(0.1371)
a
0.9473
(0.1375)
a
0.2574
(0.0311)
Years of Education
0.2566
(0.0308)
Ln(Population)
0.0684
(0.0408)
a
a
a
c
Ln(Employment)
0.0576
(0.0398)
a
Constant
c
a
a
6.3251
(0.4598)
3.5761
(0.9372)
3.7959
(0.8977)
107
104
103
50%
63%
63%
30
Observations
2
Adjusted R
Regional GDP Per Capita and Geography
Table 4 – Panel B
(1)
(2)
(3)
Temperature
‐0.0156
(0.0082)
c
‐0.0140
(0.0084)
c
‐0.0206
(0.0105)
Inverse Distance to Coast
1.0318
(0.2078)
a
0.4979
(0.1438)
a
0.5096
(0.1745)
Ln(Oil)
0.1651
(0.0477)
a
0.1752
(0.0578)
a
0.1941
(0.0440)
a
0.2751
(0.0271)
Years of Education
0.2755
(0.0171)
Ln(Population)
0.0125
(0.0168)
a
a
a
a
Ln(Employment)
Constant
c
0.0661
(0.0244)
a
a
a
8.0947
(0.2282)
6.3886
(0.1944)
5.9154
(0.2516)
1,545
107
1,478
104
833
49
2
8%
42%
50%
2
47%
60%
70%
2
34%
Yes
62%
Yes
70%
Yes
Observations
Number of countries
R Within
R Between
R Overall
Fixed Effects
31
National GDP per capita and Institutions (Table 5:A)
(1)
(2)
(3)
(4)
(5)
(6)
Years of Education
0.2566a
(0.0308)
0.2310a
(0.0344)
0.1890a
(0.0310)
0.2339a
(0.0316)
0.2291a
(0.0336)
0.2301a
(0.0350)
0.2264a 0.2355a 0.1749b
(0.0344) (0.0332) (0.0703)
Ln(Population)
0.0684
(0.0408)
c
‐0.0022
(0.0494)
0.0887
(0.0582)
0.0428
(0.0488)
0.0320
(0.0481)
0.0067
(0.0519)
0.0299
0.0611 ‐0.0782
(0.0473) (0.0457) (0.1074)
Temperature
‐0.0189
(0.0106)
c
‐0.0105
(0.0128)
‐0.0276
(0.0128)
b
‐0.0083
(0.0119)
‐0.0094
(0.0114)
‐0.0066
(0.0112)
‐0.0082 ‐0.0129 ‐0.0147
(0.0110) (0.0117) (0.0306)
Inverse Distance to Coast
2.9647
(0.5736)
a
2.3086
(0.6321)
a
2.1692
(0.7006)
a
2.5170
(0.5698)
a
2.2652
(0.5856)
a
2.2826
(0.5406)
a
2.1892
2.3979
0.2385
(0.5562) (0.5616) (2.1131)
Ln(Oil)
0.9503
(0.1371)
a
1.6367
(0.5966)
a
0.5257
(0.5050)
1.1319
(0.3309)
a
1.1739
(0.3219)
a
1.1916
(0.3302)
a
1.1165
1.2054
0.5201
(0.2950) (0.4982) (0.4921)
Informal Payments
(7)
(8)
(9)
a
a
a
b
‐0.0121
(0.0499)
a
ln(Tax Days)
‐0.5497
(0.1446)
Ln(Days without electricity)
‐0.1375
(0.0847)
Security costs
‐0.0332
(0.0250)
Access to land
‐0.7493
(0.5783)
Access to finance
‐0.5164
(0.4202)
Government Predictability
0.3835
(0.4431)
Doing Business Percentile Rank
0.6704
(1.6413)
a
Constant
3.5761
(0.9372)
Observations
a
5.1927
(1.1015)
a
5.1619
(1.2918)
a
4.6815
(0.9542)
a
4.7382
(1.0046)
a
5.1545
(0.9971)
a
a
b
4.9498
3.9328 8.6509
(1.0246) (0.9724) (3.1636)
104
73
55
75
76
80
81
72
17
63%
73%
76%
69%
69%
70%
70%
71%
34%
2 63%
73%
69%
69%
69%
69%
69%
71%
39%
2
50%
53%
60%
49%
50%
52%
52%
50%
26%
2
Adjusted R
Adj. R without institution
Adj. R without education
32
National GDP per capita, Infrastructure and Culture (T 5:B)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Years of Education
0.2566
(0.0308)
a
0.2379
(0.0338)
a
0.2642
(0.0325)
a
0.1935
(0.0498)
a
0.1818
(0.0538)
a
0.2534
(0.0347)
a
0.2394
(0.0377)
Ln(Population)
0.0684
(0.0408)
c
0.0688
(0.0414)
c
0.0653
(0.0407)
0.1238
(0.0788)
0.2169
(0.1017)
b
0.0999
(0.0640)
0.0807
(0.0450)
Temperature
‐0.0189
(0.0106)
c
‐0.0145
(0.0109)
‐0.0191
(0.0108)
c
‐0.0283
(0.0135)
b
‐0.0434
(0.0148)
a
‐0.0188
(0.0107)
c
‐0.0163
(0.0108)
Inverse Distance to Coast
2.9647
(0.5736)
a
2.7218
(0.6025)
a
3.0968
(0.6268)
a
3.6522
(0.7902)
a
4.3386
(1.0486)
a
2.7758
(0.6473)
a
2.7448
(0.5853)
Ln(Oil)
0.9503
(0.1371)
a
1.0157
(0.1438)
a
0.8737
(0.1467)
a
0.9902
(0.3207)
a
0.9751
(0.2895)
a
0.9538
(0.1443)
a
0.8792
(0.1657)
Ln(Power line density)
a
a
0.0825
(0.0934)
Trust in others
1.2472
(0.8796)
Civic values
0.4180
(0.3105)
Ln(Number of ethnic groups)
‐0.0996
(0.1550)
Probability of same language
0.4195
(0.3391)
a
3.5761
(0.9372)
Observations
c
0.1480
(0.1099)
Ln(Travel time)
Constant
a
a
3.6383
(0.9251)
b
3.0050
(1.2448)
2.3962
(2.0122)
‐0.1572
(3.2084)
a
3.4625
(0.9289)
a
3.3864
(0.9548)
104
104
104
67
57
104
103
63%
63%
63%
49%
47%
63%
62%
2 63%
63%
63%
48%
45%
63%
62%
2
50%
54%
50%
44%
42%
51%
52%
2
Adjusted R
Adj. R without infrastructure or Adj. R without education
33
Regional GDP per capita and Institutions (Table 6:A)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Years of Education in the Region
0.2758
(0.0172)
a
0.3056
(0.0298)
a
0.3620
(0.0288)
a
0.3439
(0.0481)
a
0.3343
(0.0310)
a
0.3267
(0.0218)
a
0.3273
(0.0215)
a
0.3166
(0.0207)
a
0.4141
(0.0229)
Ln(Population in the Region)
0.0126
(0.0168)
‐0.0185
(0.0495)
‐0.0175
(0.0536)
‐0.0442
(0.0613)
‐0.0191
(0.0432)
‐0.0087
(0.0316)
‐0.0098
(0.0312)
‐0.0113
(0.0305)
‐0.0026
(0.0229)
Temperature
‐0.0140
(0.0084)
c
‐0.0101
(0.0096)
‐0.0086
(0.0078)
‐0.0015
(0.0122)
‐0.0064
(0.0093)
‐0.0093
(0.0086)
‐0.0106
(0.0086)
‐0.0131
(0.0081)
0.0016
(0.0059)
Inverse Distance to Coast
0.4971
(0.1441)
a
0.4647
(0.3293)
0.8290
(0.4273)
c
0.1810
(0.4312)
0.2703
(0.3041)
0.4054
(0.2636)
0.5133
(0.2822)
c
0.4420
(0.2788)
0.0913
(0.3460)
Ln(Oil)
0.1752
(0.0578)
a
‐0.0578
(0.1283)
0.1555
(0.1319)
‐0.0584
(0.2503)
‐0.0473
(0.0862)
‐0.0224
(0.1081)
‐0.0040
(0.1113)
‐0.0170
(0.0735)
0.1834
(0.1160)
Informal Payments
‐0.0089
(0.0353)
ln(Tax Days)
‐0.0479
(0.0630)
Ln(Days without electricity)
0.0001
(0.0764)
Security costs
‐0.0004
(0.0060)
Access to land
‐0.1900
(0.1457)
Access to finance
‐0.0935
(0.1536)
Government Predictability
‐0.1251
(0.1426)
c
Doing Business Percentile Rank
‐0.6199
(0.3437)
a
Constant
a
a
a
a
a
a
a
a
a
6.3853
(0.1947)
6.5073
(0.7043)
5.7640
(0.8220)
6.8622
(0.7867)
6.4507
(0.5993)
6.3453
(0.4664)
6.2816
(0.4827)
6.4790
(0.4629)
6.3186
(0.4428)
1,469
104
338
73
255
55
216
72
352
76
387
77
381
76
368
72
172
17
2 42%
58%
66%
59%
60%
62%
62%
63%
69%
2 60%
64%
64%
53%
58%
60%
60%
63%
39%
2 62%
59%
60%
49%
53%
55%
55%
56%
51%
42%
57%
66%
59%
60%
62%
62%
62%
67%
9%
11%
14%
10%
9%
6%
5%
7%
9%
60%
64%
63%
53%
58%
60%
60%
63%
41%
42%
25%
20%
21%
26%
35%
39%
45%
50%
Observations
Number of countries
R Within
R Between
R Overall
2 Within R without institution
2 Within R without education
2 Between R without institution
2 Between R without education
34
Partial Correlation of (Log) GDP per
capita and Years of Education – Figure 6
2
RUS
IDN
IDN
KEN
MOZ
RUS
PER
ECU ARG ZAR
BRA
UKR CHN
PAN
USA
MNG
LVA THA CHN
IDN
IND
RUS
LVA IND
NAM
MEX BEN
RUS
MNG
CHN
SVK
ARG
CHL
ARE
BOL RUS EST
SRB
KHM
HND
CHE LKA
PAN
GTM
NOR
CZE
BEL
MYS
ROM
IND
RUS
ARG
GHA
BRA
MEX
TZA
CHN
ARG
MKD
JPN
PER
THA
RUS
PER
KEN
NAM
NER
UZB
LAO
ARG
LSO
ARG
HND
VNM
HUN
NIC
RUS
MEX
ZMB
COL
KGZ
BRA
RUS
VEN
ZAF
DEU
RUS
COL
HND
CHE
RUS
LTU
DOM
LVA
VEN
NGA
HRV
KGZ
IND
URYCOL
ZAR
TUR
RUS
PRY
POL
CHN
ZAR
MEX
SRB
LVA
SLV
CAN
FRA
BRA
PHL
TUR
DNK
CHN
KAZ
GBR
SVN
UKR
GRC
MDG
USA
HRV
MYS
HND
BRA
CMR
RUS
ZWE
PER
PER
RUS
CAN
ZAF
MDA
PER
COL
RUS
UKR
MYS
MEX
BRA
MEX
BRA
RUS
CHL
MAR
BLZ
PER
UKR
RUS
UGA
EGY
HRV
RUS
COLDEU
RUS
IND
BEN
PRY
ZWE
MEX
BGR
PRY
ARE
COL
SWE
LBN
CHN
ECU
NZL
NAM
USA
BRA
AUT
COL
PER
ZAR
NZL
CHE
COL
GHA
SRB
JOR
BIH
ARG
GEO
ECU
UKR
DNK
MEX
PRT
MAR
TUR
ITA
SRB
RUS
EST
COL
NER
IND
LBN
NAM
JPN
VEN
CMR ZMB
ARM
LVA
CHN
IND
MEX
SWZ
CHL
HRV
IDN NAM
USA
TZA
NER
SLV
RUS
LSO
CHE
RUS
RUS
ESP
MAR
CHN
NLD
ITA
AUT
KGZ
TUR
NPL
TZA
IND
TZA
LVA
ITA
LVA
ZWE
IND
MEX
RUS
GHA
ESP
KEN
TZA
RUS
USA
PHL
BOL
IND
IDN
BEL
PRY
MKD
MEX
LVA
GRC
DEU
MOZ
ESP
NGA
BFA
LTU
NLD
VEN
RUS
MEX
CHL
TUR
FIN
IDN
ITA
URY
VEN
NIC
NZL
NIC
ITA
SRB
GEO
GHA
AUS
MNG
DEU
PRT
GAB
MNG
NLD
AZE
LAO
TZA
PHL
KAZ
ARG
ECU
KHM
NOR
SEN
IRL
LSO
TZA
ZAF
RUS
AUS
USA
USA
CHE
VEN
ESP
KGZ
BEN
ARG
FIN
GHA
JPN
MKD
NGA
SEN
ARG
URY
RUS
BRA
ITA
DOM
CMR
JPN
EST
POL
USA
COL
RUS
CHL
PRY
SVN
UKR
URY
DEU
GRC
POL
JPN
ARG
KHM
NIC
IND
ZWE
SYR
MWI
NOR
MYS
HUN
LTU
LAO
IDN
LVA
ITA
VEN
BRA
CHN
MYS
MKD
LVA
JPN
ITA
VEN
ZMB
GBR
ISR
MNG
NIC
ARG
CAN
IDN
NIC
SYR
SYR
MNG
IDN
UZB
JPN
PHL
VNM
ECU
BFA
SLV
POL
URY
RUS
UGA
BLZ
JPN
SVK
UKR
ECU
BEL
LAO
MNG
ROM
UKR
RUS
COL
ECU
LAO
SEN
TZA
MNG
JOR
JPN
ESP
LVA
SRB
JPN
SEN
EGY
IND
ISR
HRV
LAO
PRY
SRB
USA
VEN
PHL
AUS
TUR
FRA
BRA
BEN
NAM
VNM
ARM
RUS
HUN
URY
GEO
GAB
ECU
BFA
RUS
SYR
HND
RUS
TZA
NIC
NZL
DNK
COL
JPN
USA
SVN
ARM
ZAR
USA
CHN
MEX
ITA
SLV
COL
POL
ESP
PAK
BFA
USA
PHL
JPN
NLD
MEX
HND
SYR
HRV
USA
JOR
NIC
GTM
MNG
ARM
MOZ
BOL
BEL
URY
UKR
CHE
PAK
JOR
JPN
PRY
LAO
BFA
ZAF
PHL
ITA
AUT
GRC
KAZ
CHN
CMR
CHL
NAM
EST
DNK
MDG
COL
RUS
BFA
NLD
RUS
USA
TZA
NOR
IDN
NOR
COL
ZAR
PRY
TZA
DEU
LTU
LAO
SVN
FRA
LVA
DOM
ISR
IDN
GEO
URY
DNK
PRT
BRA
RUS
JPN
GTM
JOR
NER
EST
GAB
DOM
PRY
USA
MNG
KHM
FRA
BFA
SYR
MNG
JPN
LTU
PHL
VEN
LVA
DOM
USA
COL
CHE
LAO
SEN
DNK
LVA
JPN
POL
JPN
NPL
PHL
SLV
JPN
COL
USA
USA
CAN
VEN
FRA
COL
UKR
MYS
URY
LSO
NOR
CHE
ROM
GRC
MYS
EST
BFA
JPN
EST
CHE
PER
HRV
SVN
JPN
POL
GHA
SWZ
DEU
PAK
AZE
MEX
GBR
URY
TZA
CAN
GBR
SLV
FRA
BEL
HND
SWE
BLZ
ARM
ARG
PER
URY
BLZ
BGR
AUT
FRA
JPN
NZL
USA
ITA
ZMB
LAO
MAR
IND
USA
JOR
POL
USA
JPN
ESP
SRB
ECU
NZL
AZE
FRA
AUT
MEX
GBR
DNK
FRA
CAN
VEN
LVA
ECU
VEN
PAN
NPL
SYR
JPN
KHM
GEO
SRB
IDN
GEO
GEO
NOR
NOR
PER
FRA
IND
AZE
BRA
KEN
MAR
PAN
SRB
UKR
JPNPER
PHL
AZE
LTU
NIC
VNM
AUS
GRC
BFA
BEN
URY
JPN
GBR
BOL
SYR
SVN
SYR
IND
FRA
KHM
SWE
PHL
POL
USA
GRC
NOR
BEL
SWE
NGA
AUS
CHE
AZE
SVN
MDG
ESP
HND
MYS
SRB
DNK
ROM
CHE
FRA
CZE
NZL
LKA
USA
NZL
SYR
BFA
EST
ZAF
VEN
USA
CHE
MAR
JOR
SEN
ECU
JPN
BGR
CZE
RUS
AZE
GHA
NLD
ISR
ECU
LVA
DNK
DEU
GTM
SLV
RUS
BRA
POL
CHN
LVA
HRV
SVK
MWI
RUS
ESP
CAN
SLV
RUS
GRC
DEU
SEN
JPN
JPN
TUR
JPN
SVK
USA
SRB
PER
BGR
USA
HRV
IDN
DEU
ARM
ECU
NLD
SWE
MAR
HND
CHN
CHL
IND
JPN
HND
SWZ
LAO
ESP
SYR
JPN
GEO
TZA
ESP
GBR
UGA
FRA
NOR
PAN
USA
FRA
GTM
HRV
VNM
NOR
FIN
GRC
NZL
CZE
MNG
URY
JOR
DNK
PRY
LVA
NIC
BEN
ZAR
DOM
GBR
SVN
PRT
POL
MAR
KHM
JPN
CMR
EST
AUT
SLV
USA
FRA
GTM
HRV
ARM
BEN
MDG
JOR
UKR
USA
NLD
FRA
JPN
SWE
SWE
PER
IND
MEX
MAR
BLZ
NOR
BGR
ESP
CHE
LKA
DEU
JPN
ZWE
HND
CHL
USA
IND
LSO
NOR
NOR
LVA
MNG
KHM
ARM
NZL
USA
VNM
USA
FIN
BOL
ISR
LBN
NIC
NIC
NER
LTU
JPN
EST
ESP
SVN
DNK
MNG
IDN
SVK
JPN
RUS
MDA
NPL
CHE
LKA
LSO
NLD
LAO
RUS
URY
LAO
FRA
HRV
AZE
CZE
CHE
RUS
ECU
CMR
CHN
MWI
ARE
MAR
HRV
IND
ROM
SEN
CZE
MKD
KHM
FRA
SLV
PRY
FRA
MOZ
HRV
BEL
AUS
UKR
KHM
IDN
TZA
IDN
ZAR
LBN
PRT
LVA
GEO
UKR
GEO
MEX
NOR
LVA
ARM
NOR
ESP
USA
SVK
RUS
CHE
CHE
KAZ
ZMB
PER
NZL
HRV
BEN
MDA
LSO
MEX
GBR
DOM
AUS
TUR
RUS
ITA
CMR PERMYS
JOR
LVA
GRC
MAR
LKA
CHE
JPN
GBR
ZAF
HND
URY
ZAF
MAR
JPN
MAR
NOR
NLD
CHL
HRV
UKR
CAN
JOR
LAO
BRA
CMR
UKR
ZMB
SLV
SVN
ZWE
CHE
MEX
ZMB
HRV
POL
PHL
KGZ
MOZ
USA
VEN
ZWE
UZB
JPN
DNK
EST
CAN
RUS
BFA
USA
BIH
SLV
BEN
VEN
SYR
PRY
PRT
NZL
URY
KHM
DNK
PER
PHL
ARE
ESP
LAO
NOR
ISR
MYS
IND
ZWE
POL
USA
SYR
USA
PHL
CZE
ROM
LVA
UKR
GRC
LTU
AUT
BIH
COL
COL
JPN
SRB
IND
NLD
PAK
GBR
LKA
RUS
RUS
MDG
SLV
JOR
MKD
SYR
TZA
LVA
JPN
HUN
SEN
POL
CHN
VEN
HRV
ITA
KAZ
RUS
THA
AUT
IND
VEN
JPN
BRA
MOZ
ARE
ARG
NPL
CZE
EST
CMR
IND
BGR
MNG
IRL
LVA
SVN
ECU
JPN
MAR
USA
MDA
USA
CHN
LBN
BEL
LKA
CHE
HUN
MYS
ESP
BFA
LVA
EST
PHL
BRA
GRC
ROM
IND
CAN
IDN
MEX
AUS
PHL
CHN
HND
USA
TZA
ITA
NOR
NAM
LSO
UKR
PRT
RUS
IDN
UZB
TUR
ZMB
DNK
NZL
KEN
HUN
MDG
URY
KHM
ARM
ARG
NLD
SRB
DNK
RUS
FIN
ARG
USA
LBN
IDN
BRA
RUS
CHE
AUT
CHE
IND
SVN
COL
MEX
SWZ
MOZ
USA
LSO
NIC LAO
ARG
NIC
MNG
NIC
LTU
CAN
SVK
EGY
CHN
UKR
ITA
POL
BEL
DEU
COL
EGY
ECU
SEN
USA
NZL
BEN
HUN
UKR
RUS
PRY
CHE
MOZ
MNG
CHN
BEL
RUS
VNM
CHL
BOL
GRC
MOZ
BOL
DEU
GHA
SLV
ZWE
CHN
CHN
JPNDEU
UKR
JPN
DEU
SRB
NER
IDN
PRY
RUS
HND
LAO
VEN
MEX
EST
DEU
HND
BEL
TZA
ECU
GTM
VEN
USA
CHN
CAN
NAM
RUS
NIC
BFA
IND
CHN
LVA
UKR
URY
SRB
DOM
ECU
TZA
CMR PER
PRY
COL
LSO
ITA
IDN
BRA
ESP
DOM
HRV
ROM
RUS
BRA
KHM
KGZ
LVA
LAO
ITA
PRY
ITA
NIC
GEO
ITA
EST
COL
GAB
CHL
GTM
KEN
IDN
MNG
HRV
ZMB
RUS
HND
MEX
USA
BEN
MYS
CHN
VEN
IND
ZWE
SRB
IND
RUS
UZB
IND
BRA
RUS
LVA
NER
PRY
ECU
RUS
MEX
VEN
TUR
NAM
CHN
TZA
IDNNGA
ECU
TZA
CHN
VNM
HND
CHN
MEX
COL
MEX
RUS
RUS
KAZ
MEX
PAN
MYS
PER
SRB
SVK
CHL
MEX
RUS
LVA
ZAF
SRB
VEN
UKR
PER
TZA
MNG
UGA
MKD
COL
RUS
PAN
KEN
UKR
URY
COL
PAN
BLZ
NAM
NER
MKD
ARG
ARG
PAN NGA PHL
BOL
BRA
IDN
COL
TUR
LTU
RUS
IND
ZAF
MNG
RUS
KGZ
BEN
CHN
MOZ
ECU
THA
BRA
BRA
LVA
COL
COL
KGZ
RUS
CHN
ARG
HND
ARE
PER
LVA
IND
TUR
RUS
RUS
MNG
RUS
MEX
CHL
NAM
BRALVA MYS
GHA
COL
GHA
PER
NAM
PER
ARG
MEX
ARG
RUS
RUS
RUS
BRA
ZAR
COL CHN
ZAR
RUS
IDN
IDN
PER
IDN
THA
ZAR
ARG
INDARG
RUS
KEN
Ln(Regional GDP per capita
-1
0
1
COL
-2
RUS
-4
-2
0
Years of Education
2
4
coef = .28573483, (robust) se = .01343762, t = 21.26
35
Regional GDP per capita, Infrastructure and Culture (T 6:B)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Years of Education in the Region
0.2758
(0.0172)
a
0.2713
(0.0187)
a
0.2627
(0.0197)
a
0.3021
(0.0286)
a
0.2986
(0.0305)
a
0.2644
(0.0181)
a
0.2719
(0.0175)
Ln(Population in the Region)
0.0126
(0.0168)
0.0101
(0.0168)
0.0023
(0.0184)
0.0091
(0.0187)
0.0138
(0.0193)
0.0170
(0.0173)
0.0115
(0.0157)
Temperature
‐0.0140
(0.0084)
c
‐0.0142
(0.0085)
c
‐0.0166
(0.0085)
c
‐0.0015
(0.0060)
‐0.0038
(0.0056)
‐0.0154
(0.0090)
c
‐0.0140
(0.0080)
Inverse Distance to Coast
0.4971
(0.1441)
a
0.4872
(0.1427)
a
0.4626
(0.1438)
a
0.4750
(0.2590)
c
0.4093
(0.2713)
0.4351
(0.1358)
a
0.5162
(0.1450)
Ln(Oil)
0.1752
(0.0578)
a
0.1793
(0.0584)
a
0.1864
(0.0582)
a
0.0534
(0.0669)
0.0354
(0.0572)
0.1922
(0.0613)
a
0.1772
(0.0591)
Ln(Power line density)
c
a
a
0.0199
(0.0198)
c
Ln(Travel time)
‐0.0456
(0.0231)
Trust in others
‐0.0611
(0.0868)
Civic values
‐0.0040
(0.0231)
b
Ln(Number of ethnic groups)
‐0.0504
(0.0249)
Probability of same language
0.1723
(0.2067)
a
Constant
a
a
a
a
a
a
a
6.3853
(0.1947)
6.4350
(0.1928)
6.9287
(0.3351)
6.0940
(0.2863)
6.0196
(0.3245)
6.5272
(0.1679)
6.2956
(0.2337)
1,469
104
1,469
104
1,469
104
699
65
635
70
1,468
104
1,445
103
2 42%
42%
43%
49%
48%
42%
42%
2 60%
60%
60%
50%
50%
60%
60%
R Overall
2 62%
62%
61%
50%
47%
62%
62%
2 42%
42%
42%
49%
48%
42%
42%
2 10%
13%
16%
11%
11%
14%
11%
2 60%
60%
60%
51%
50%
60%
59%
2 41%
Yes
51%
Yes
47%
Yes
7%
Yes
16%
Yes
47%
Yes
49%
Yes
Observations
Number of countries
R Within
R Between
Within R without institution
Within R without education
Between R without institution
Between R without education
Fixed Effects
36
National GDP Per Capita and Standard
Measures of Institutions (Table 7)
(1)
(2)
(3)
(4)
(5)
(6)
Years of Education
0.2567
(0.0308)
a
0.2200
(0.0433)
a
0.2069
(0.0438)
a
0.1626
(0.0480)
a
0.2448
(0.0363)
a
0.1850
(0.0351)
a
Ln(Population)
0.0683
(0.0410)
c
0.0354
(0.0487)
0.0559
(0.0470)
‐0.0356
(0.0482)
0.0732
(0.0533)
0.0504
(0.0370)
Temperature
‐0.0189
(0.0106)
c
‐0.0179
(0.0118)
‐0.0135
(0.0109)
0.0024
(0.0106)
‐0.0181
(0.0126)
‐0.0100
(0.0104)
Inverse Distance to Coast
2.9646
(0.5742)
a
2.3421
(0.7800)
a
2.3853
(0.6050)
a
2.3974
(0.5941)
a
2.9603
(0.6208)
a
1.9906
(0.5463)
Ln(Oil)
0.9503
(0.1373)
a
0.7877
(0.4564)
c
1.0708
(0.1729)
a
0.8965
(0.1100)
a
1.0720
(0.4094)
b
0.9928
(0.2013)
a
a
a
Autocracy
‐0.5994
(0.2184)
b
Executive Constraints
0.1633
(0.0696)
a
Expropriation Risk
0.3952
(0.0986)
c
Proportional Representation
0.3972
(0.2328)
a
Corruption
0.2130
(0.0479)
a
Constant
3.5771
(0.9416)
Observations
a
5.3781
(1.3861)
a
3.7896
(1.0059)
b
3.1830
(1.3630)
a
3.2958
(1.0503)
a
4.1183
(0.8118)
103
80
101
81
97
103
63%
67%
65%
70%
63%
69%
2 63%
64%
63%
63%
62%
63%
2
50%
60%
59%
67%
52%
63%
2
Adjusted R
Adj. R without institution
Adj. R without education
37
Firm Level Productivity and Regional Education (Table 8)
Dependent Variable:
Logarithm of Sales per employee
Logarithm of Wages per employee
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(1)
Years of Education in the Region
0.0655
(0.0202)
a
0.0639
(0.0185)
a
0.0954
(0.0280)
a
0.0950
(0.0279)
a
0.0580
(0.0162)
a
0.0577
(0.0159)
a
0.0840
(0.0233)
a
0.0843
(0.0234)
Ln(Population in the Region)
0.0920
(0.0321)
a
0.0803
(0.0297)
a
0.1437
(0.0501)
a
0.1409
(0.0504)
a
0.0682
(0.0425)
0.0622
(0.0418)
0.0135
(0.0352)
0.0159
(0.0354)
Years of Education of managers
0.0534
(0.0047)
a
0.0352
(0.0048)
a
0.0257
(0.0062)
a
0.0243
(0.0057)
a
0.0315
(0.0038)
a
0.0215
(0.0036)
a
0.0118
(0.0044)
a
0.0131
(0.0042)
.
.
0.1497
(0.0154)
a
.
.
0.0113
(0.0176)
.
.
0.0827
(0.0150)
a
.
.
‐0.0095
(0.0108)
a
0.0384
(0.0056)
0.0378
(0.0058)
0.0195
(0.0036)
a
0.0151
(0.0036)
a
0.0146
(0.0033)
0.0152
(0.0033)
a
.
.
0.2248
(0.0173)
a
0.2232
(0.0172)
a
.
.
.
.
a
.
.
.
.
0.1787
(0.0086)
a
0.1794
(0.0089)
a
a
Ln(Employees)
Years of Education of workers
0.0349
(0.0053)
a
0.0279
(0.0054)
Ln(Expenditure on energy/employee)
0.3577
(0.0185)
a
0.3554
(0.0177)
.
.
.
.
.
.
0.3258
(0.0132)
a
0.3250
(0.0136)
a
Ln(Property, Plant, Equipment / employees)
a
Constant
Observations
Number of Countries
2 Within R
2 Between R
2 Overall R
a
a
a
a
a
a
a
a
a
a
a
5.1202
(0.3706)
5.0055
(0.3373)
4.8529
(1.1885)
4.8850
(1.1887)
5.1007
(0.5225)
5.0322
(0.5199)
6.6732
(0.7223)
6.6461
(0.7248)
13,248
29
13,248
29
19,305
22
19,305
22
12,782
27
12,782
27
19,209
22
19,209
22
30%
32%
31%
31%
20%
21%
13%
38 13%
90%
90%
59%
59%
88%
87%
57%
57%
74%
74%
54%
54%
69%
68%
44%
44%
Interpretation of the coefficients (1)

Columns (1)-(4) of Table 8  coefficient on SW ≈0.035  α*μW ≈ 0.35
 Standard assumption: α=0.6  μW= 5.8%.
 Alternative assumption: α=0.55  μW= 6.4%.

Table 8 also implies an overall capital share roughly equal to 0.35  δ + β =0.35.
 Entrepreneurial rents: (1–α–β–δ) = (1 – 0.55 – 0.35) = 0.1.
 Since the estimated coefficient on SE = (1– α–β–δ)*μE = 0.035  μE = 35%.
 Entrepreneurial inputs are a neglected but critical channel through which schooling
and human capital affect productivity .

The coefficient on population in Table 8 is roughly equal to 0.09  γ≈ 0.09.
 Consistent with the regional regressions in Table 4  the coefficient on population
is roughly equal to 0.01  γ – β/(1 – δ)= 0.01.
 Given that γ = 0.09 and β + δ = 0.35, this condition yields β ≈ 0.05.
 β = 0.05 and γ = 0.09 are roughly consistent with our results.
39
Interpretation of the coefficients (2)

The coefficient on regional schooling in Table 8 roughly equal to 0.065  γ*ψ*μ≈0.065.
 γ = 0.09  that ψ* μ=0.72.
 The coefficient on schooling in the regional regressions equals (1+ γψ – β/(1-δ)) μ.
 Table 4  (1+ γψ – β /(1-δ)) μ ≈0.27.
 Given γ = 0.09 and β /(1-δ) = 0.08, we are left with 2 equations with 2 unknowns:
1. (1+ 0.09ψ – 0.08) μ ≈ 0.27 ;and
2. ψ*μ ≈ 0.72.
 μ ≈ 0.22 and ψ ≈3.27.
 Estimates imply a large effect of schooling on productivity via social
interactions.

The above parameter values and at a reasonable share of housing consumption of
θ = 0.4 the spatial equilibrium is stable, since
(β – ψγ)(1 – θ) + θ(1 – δ) = – (0.21)(0.6) + (0.4)(0.73) > 0.
40
Firm Level Productivity and Regional Education (Table 8A)
Dependent Variable:
Logarithm of Sales per employee
Logarithm of Wages per employee
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(1)
Temperature
0.0102
(0.0129)
0.0123
(0.0118)
0.0171
(0.0087)
c
0.0171
(0.0087)
c
‐0.0117
(0.0115)
‐0.0104
(0.0110)
0.0149
(0.0051)
a
0.0149
(0.0051)
Inverse Distance to Coast
0.2755
(0.3492)
0.1574
(0.3450)
0.8401
(0.2565)
a
0.8380
(0.2565)
a
‐0.1088
(0.3266)
‐0.1830
(0.3294)
0.0437
(0.2084)
0.0530
(0.2095)
Ln(Oil)
‐0.8864
(0.2793)
a
‐0.7033
(0.3070)
b
0.2600
(0.5745)
0.2639
(0.5723)
‐0.6647
(0.3179)
b
‐0.5661
(0.3380)
c
0.1110
(0.3702)
0.0945
(0.3681)
Ln(Power line density)
‐0.0246
(0.0317)
‐0.0266
(0.0313)
0.1157
(0.0430)
a
0.1155
(0.0429)
a
‐0.0125
(0.0333)
‐0.0159
(0.0334)
0.0285
(0.0412)
0.0293
(0.0412)
Access to finance
‐0.0747
(0.0772)
‐0.0727
(0.0715)
‐0.0635
(0.0989)
‐0.0633
(0.0990)
‐0.1051
(0.0877)
‐0.1059
(0.0839)
‐0.1238
(0.0778)
‐0.1248
(0.0783)
Years of Education in the Region
0.0685
(0.0262)
b
0.0721
(0.0247)
a
‐0.0203
(0.0321)
‐0.0202
(0.0321)
0.0666
(0.0267)
b
0.0708
(0.0264)
a
0.0602
(0.0298)
b
0.0599
(0.0298)
Ln(Population in the Region)
0.1046
(0.0344)
a
0.0906
(0.0316)
a
0.0454
(0.0381)
0.0448
(0.0386)
0.0758
(0.0428)
c
0.0692
(0.0422)
‐0.0353
(0.0289)
‐0.0327
(0.0289)
Years of Education of managers
0.0531
(0.0047)
a
0.0351
(0.0047)
a
0.0274
(0.0061)
a
0.0270
(0.0054)
a
0.0315
(0.0037)
a
0.0216
(0.0036)
a
0.0129
(0.0043)
a
0.0145
(0.0040)
.
.
0.1486
(0.0153)
a
.
.
0.0030
(0.0175)
.
.
0.0821
(0.0150)
a
.
.
‐0.0129
(0.0109)
a
0.0410
(0.0059)
0.0408
(0.0061)
0.0193
(0.0037)
a
0.0151
(0.0037)
a
0.0158
(0.0035)
0.0165
(0.0034)
a
.
.
0.2250
(0.0173)
a
0.2235
(0.0172)
a
.
.
.
.
a
.
.
.
.
0.1759
(0.0084)
a
0.1768
(0.0088)
a
a
Ln(Employees)
Years of Education of workers
0.0344
(0.0053)
a
0.0276
(0.0054)
Ln(Expenditure on energy/employee)
0.3574
(0.0184)
a
0.3552
(0.0176)
.
.
.
.
.
.
0.3193
(0.0123)
a
0.3190
(0.0127)
a
Ln(Property, Plant, Equipment / employees)
a
Constant
Observations
Number of Countries
2 Within R
2 Between R
2 Overall R
a
a
a
a
a
a
a
b
a
a
a
a
4.9971
(0.5219)
5.0412
(0.4805)
5.7485
(1.0095)
5.7554
(1.0110)
6.1865
(0.7436)
6.1040
(0.7271)
7.2040
(0.6329)
7.1742
(0.6343)
13,248
26
13,248
26
19,305
19
19,305
19
12,782
24
12,782
24
19,209
19
19,209
19
30%
32%
28%
29%
19%
20%
16%
16%
91%
91%
48%
47%
81%
79%
47%
46%
69%
67%
45%
42%
61%
58%
44%
42%
41
Firm Size and Regional Employment

Productive regions employ more human capital (i.e., HG > HB) and such human capital is of
better quality, namely EG(h) > EB(h).
h



hm
h
The skill distribution in the unproductive region is truncated at hm.
The skill distribution in the productive region jumps to the red line for h > hm.
Proposition 1: If p is sufficiently large, there are 2 thresholds z1 and z2 such that for
 AG /
1-
AB (  - )(1- ) (1- )

 ( z1, z2 )
42
productive regions have larger: i) average firm, ii) share of workers in the population.
Regional GDP, Education, Size of Establishments,
and Labor participation (Table 9)
Dependent Variable:
Ln(Establishments/Population) Ln(Employees/Establishments) Ln(Employees/Population) Ln(Employees Big Firms/Employees)
Years of Education in the Region
0.2967
(0.0314)
a
0.1233
(0.0227)
a
0.3418
(0.0273)
a
0.2445
(0.0374)
Constant
‐5.8626
(0.2571)
a
0.8855
(0.2093)
a
‐4.3992
(0.2119)
a
‐3.6568
(0.4299)
951
92%
Yes
983
83%
Yes
988
94%
Yes
501
95%
Yes
Observations
2
Adjusted R
Country Fixed Effects
a
a
43
PHL SEN
Firms/Pop
-2
0
2
BFABEN
CHN
PHL
TZACHN
DNK ITABFA
RUS
ITAKAZ CHN
ITA
PHLDEU
ITA
MOZ
UKR
RUS
ITA
BEN
CMR
JOR
MDG
CHN
UGA
LVA
TUR
MOZ
HRV
SEN
KHM
RUS
ITA
BRA
MNG
RUS
CHN
ITA
BRA
ITA
JOR
GEO
RUS
ESP
CMR
MOZ
TZA
HRV
PER
JOR
GRC
ITA
RUS
KHM
BRA
SYR
KHM
BRA
PRY
LAO
LAO
DNK
GHA
SYR
TZA
EST
SYR
LAO
KHM
RUS
CHN
RUS
HRV
ROM
IND
BRA
TZA
TZA
LKA
PER
CHE
CHN
LVA
LTU
MNG
BFA
NPL
PHL
DNK
DNK
GEO
USA
PER
THA
IND
BEL
ESP
KHM
USA
MNG
POL
KHM
IND
CAN
BEN
HRV
LVA
MDG
KHM
ESP
GHA
IND
KHM
BRA
MYS
MKD
BRA
RUS
ESP
LAO
LVA
MYS
MNG
CHN
GHA
CMR
LAO
RUS
ESP
USA
HUN
LVA
IND
POL
HRV
AUT
MYS
BEN
KHM
PER
LVA
UKR
MNG
SLV
IND
POL
RUS
RUS
UKR
KHM
LVA
POL
USA
TUR
CMR
MYS
MNG
ROM
BRA
SLV
RUS
UKR
MYS
SWE
ITA
EST
RUS
USA
PAN
FIN
NPL
LVA
BRA
CAN
LTU
GRC
GHA
LVA
VNM
FRA
ESP
NLD
TZA
USA
USA
RUS
ROM
PAN
VNM
IND
BRA
IND
KHM
MDG
HRV
MOZ
DEU
MOZ
BRA
SYR
RUS
CAN
IND
TZA
THA
ARE
LVA
EST
PHL
BFA
LVA
JPN
RUS
UKR
CHE
RUS
SYR
NOR
SLV
FRA
UKR
PAN
JOR
TUR
POL
LTU
BEL
PER
PAN PER
ESP
RUS
MYS
NLD
PHL
CZE
HRV
KHM
SLV
IND
EST
POL
SYR
PER
NOR
GRC
JPN
ESP
NOR
CZE
USA
ARE
USA
EST
VNM
ITA
IND
UKR
FRA
CMR
LVA
NZL
SYR
THA
VNM
LAO
BGR
RUS
IND
TUR
DNK
ROM
JPN
NPL
JPN
AUT
FRA
USA
POL
MNG
CHE
IND
MKD
LAO
NZL
JPN
PER
IND
CZE
JPN
CHN
JPN
NLD
JPN
UKR
RUS
POL
FRA
GEO
RUS
USA
LAO
BEL
TUR
HUN
CZE
SYR
PER
LAO
LAO
MKD
JPN
KAZ
MNG
EST
EST
ARE
AUT
JPN
RUS
HUN
ESP
PHL
GEO
NOR
TZA
AUT
CZE
BGR
MNG
MNG
POL
MKD
FRA
SLV
CMR
MYS
USA
JPN
CHE
NLD
LVA
JPN
FRA
HRV
RUS
USA
PER
USA
RUS
CHE
BEL
PER
DEU
BEL
IRL
NZL
JPN
MYS
FRA
SWE
SLV
LVA
CHE
LKA
LTU
UKR
LVA
JPN
SLV
RUS
TUR
IND
USA
VNM
LKA
BGR
PER
BEL
TZA
ARE
NZL
NOR
ESP
POL
FRA
JPN
BRA
USA
USA
NZL
RUS
SWE
IND
FRA
TZA
SWE
DNK
AUT
CHE
JPN
JPN
NZL
JOR
ESP
NLD
MKD
PER
JPN
HRV
IND
PER
RUS
NLD
JPN
CAN
NZL
TZA
CAN
ESP
HRV
UGA
SLV
USA
RUS
CHE
JPN
CHE
NZL
GRC
USA
LKA
RUS
HUN
CAN
NOR
JPN
NZL
RUS
MKD
FRA
BGR
RUS
RUS
MNG
LKA
EST
USA
NOR
IND
NOR
JPN
CHE
JPN
TUR
IND
GRC
SYR
GRC
DNK
NOR
LVA
NZL
JPN
IND
NOR
BEN
DEU
USA
USA
CHN
CHE
CAN
FIN
NLD
PER
IRL
ARE
NLD
POL
CAN
IND
SWE
HUN
PAN
BRA
DNK
NOR
GRC
USA
FIN
RUS
AUT
BGR
MNG
JPN
PHL
LVA
LVA
UKR
BEL
AUT
EST
JOR
JPN
NOR
SWE
FRA
GEO
USA
SLV
LTU
HRV
AUT
MNG
RUS
PHL
DEU
LTU
NOR
KHM
EST
FRA
NLD
VNM
CHE
RUS
RUS
ARE
PER
LAO
BRA
LAO
BEL
ESP
USA
SLV
POL
USA
GEO
UKR
USA
GEO
UKR
LTU
JPN
CAN
CHE
HRV
SWE
NLD
JPN
DNK
PAN PAN
GHA
LAO
GRC
CMR
UKR
MNG
GRC
BRA
CHE
CAN
GEO
IND
DNK
FIN
CHN
NOR
GHA
GHA
RUS
NZL
IND
LVA
BEL
SLV
JOR
LKA
POL
EST
CHE
BGR
UKR
KAZ
RUS
GEO
JOR
GRC
FRA
PAN
IND
GRC
HRV
FRA
MNG
MYS
VNM
EST
BRA
BEL
TZA
UKR
PER
USA
UGA
LVA
DEU
EST
LTU
ROM
CHE
NLD
PER
MNG
CHN
PER
BRA
HUN
FRA
DEU
DEU
HUN
DNK
UKR
CHN
TZA
RUS
CHN
POL
RUS
TUR
LVA
SLV
LAO
UKR
UKR
MNG
RUS
NPL
HRV
UKR
DEU
LAO
GRC
PER
IND
LTU
CHN
SLV
ROM
RUS
BRA
ARE
RUS
ESP
BFA
MKD
RUS
CHN
GHA
JPN
IND
BRA
UKR
PER
LVA
CHN
MOZ
JPN
BRA
BEL
LVA
GEO
MOZ
ROM
RUS
SYR
MYS
JOR
IND
TUR
HRV
RUS
LVA
MYS
BFA
RUS
USA
PHL
BEN
THA
CZE
JPN
PER
GHA
BRA
LKA
MNG
LAO
HRV
MNG
UKR
BEN
CHN
CZE
KHM
RUS
TUR
CHN
SYR
RUS
ROM
THA
AUT
HRV
LVA
GEO
PER
CHN
DEU
PER
IND
CHN
TZA
PAN
RUS
BRA
CMR
MNG
JOR
KAZ
MDG
CAN
GHA
BEN
HRV
MDG
BRA
LVA
RUS
CHN
CHN
KAZ
VNM
TZA
TZA
SYR
TZA
CHN
DNKMYS LAO
SYR
PHL
KHM
BEN
RUS
ITA
USA
RUS
EST
PRY
NPL
DNK
SYR
PHL
BRA
LVA
MOZ
USA
IND
TZA
USA
RUS
TZA
BRA
JOR
POL
CHN
BFA
MNG
BFA
ITA
USA
BEN
BRA
DNK
PHL
USA
MDG
TUR
BFA
BEN
IND
POL
JOR
IND
BEN
LVA
RUS
UGA
MYS
CHN
PHL
HRV
CMR
DNK
TZA
RUS
LAO
PHL
CMR
BFA
CHN
BFA
SEN
ESP PHL
ESP
ITA
MOZ
PHL
ITA
IND
-4
-2
0
2
Years of Education
4
0
2
Years of Education
4
coef = .34175848, (robust) se = .02725977, t = 12.54
IND
%Employment Large Firms
-2
-1
0
1
2
4
-4
Employees/Pop
-2
0
2
PHL
SEN
DNK
PHL
SEN
ITA
TZA
MOZ
CMR
CMR
ITA CMRGEO
ITA
THABEN
UKR
SRB
ITA
ITA
UGA
DEU
ARG
TUR
CHN
KAZ
ITA
MYS
MAR
CHN
LVA
PAN
PHLCMR
SLV
ESP
SLV
MDA
JOR
ESP
RUS
MYS
MYS
IND
LKA
MEX
ESP
BRA
ITA
ITA
DNK
THA
MOZ
CHN
ITA
NPL
MOZ
ARG
CHN
DNK
NPL
GHA
GRC
TZA
LTU
USA
IND
PAN
PER
MNG
ESP
ESP
EST
MEX
JOR
JOR
BEN
CMR
MYS
GEO
BEN
TZA
CHE
BRA
MEX
TUR
TZA
RUS
MEX
LAO
LAO
BEL
MEX
PER
LVA
IND
BRA
ARG
UKR
GHA
MYS
CHN
PER
NLD
ESP
JPN
SLV
HUN
UKR
MAR
CHN
GRC
MEX
MEX
LAO
MAR
SWE
AUT
AUT
BEL
HRV
MEX
GHA
SYR
VNM
IND
POL
CMR
PAK
MKD
VNM
LVA
ARE
IND
HRV
MNG
ESP
ROM
CHE
BEL
ARG
BRA
LVA
GRC
ESP
LAO
ITA
IND
SLV
MEX
LTU
PER
BRA
HUN
LVA
FRA
NOR
LVA
MNG
BRA
LAO
USA
DNK
MOZ
HRV
DEU
LKA
UKR
RUS
FRA
CAN
LAO
VNM
ESP
PAK
NLD
MEX
MAR
MEX
POL
SYR
PAN
JOR
TZA
GRC
ROM
ESP
FIN
TUR
ARG
BRA
LTU
MNG
BEL
RUS
GEO
ARE
CHN
UKR
PHL
FRA
LVA
RUS
LVA
CHN
UKR
CHN
RUS
LVA
LAO
IND
FRA
POL
TZA
FRA
LVA
HRV
MNG
TUR
CZE
SYR
MEX
SLV
CHE
PAN
HRV
MKD
NZL
IND
GRC
FRA
MNG
ESP
CHN
UKR
USA
POL
MEX
BRA
CHE
RUS
USA
ROM
PER
IND
NLD
LVA
EST
AUT
POL
SWE
UKR
RUS
UKR
BGR
ARG
FRA
USA
POL
ARG
EST
IND
DEU
PHL
MYS
GRC
CZE
HRV
USA
JPN
CAN
GEO
POL
SYR
LAO
RUS
NZL
PER
MKD
MYS
KAZ
MEX
EST
BGR
ARG
DEU
UKR
JPN
LVA
ARG
ROM
TZA
JPN
TUR
RUS
GHA
DNK
HRV
PER
JPN
RUS
HRV
RUS
BEN
NLD
GHA
BEN
VNM
MNG
CZE
BRA
USA
LVA
TZA
PER
TZA
RUS
RUS
TZA
RUS
LVA
ARG
RUS
RUS
JPN
USA
CZE
EST
USA
NPL
NOR
MNG
BRA
RUS
MKD
MKD
IND
MEX
CZE
NZL
FRA
HRV
IND
POL
JPN
NLD
LVA
CHE
SYR
PER
IND
RUS
GEO
NZL
IND
USA
PER
NZL
FRA
RUS
ITA
DEU
MDA
FRA
CHE
FRA
CHE
DNK
JPN
RUS
MAR
JPN
NLD
SWE
ARG
MNG
IND
IND
ARE
SYR
UKR
ESP
PER
IRL
JPN
RUS
EST
HRV
SYR
JPN
EST
LVA
POL
USA
HRV
NOR
USA
LTU
TZA
AUT
ARE
NOR
LKA
USA
CHE
TUR
USA
NOR
CAN
MEX
JPN
BGR
BEL
IND
PER
SYR
IND
PER
NOR
MAR
CAN
EST
USA
CAN
JPN
JOR
USA
BIH
JPN
HRV
FIN
CAN
JPN
LAO
PAK
POL
NZL
JPN
SWE
GRC
RUS
USA
USA
BRA
LVA
BEN
NZL
USA
RUS
BGR
SLV
RUS
NOR
RUS
JPN
NZL
RUS
IND
USA
RUS
PAN
RUS
BRA
CAN
AUT
JPN
IRL
MYS
IND
BEL
POL
RUS
HRV
CHN
JPN
MEX
MNG
RUS
NZL
USA
JPN
MNG
CHN
IND
CHE
NOR
JPN
LVA
CZE
CHE
AUT
EST
VNM
ARG
NOR
POL
PER
PER
MKD
JPN
BRA
HRV
NOR
JPN
NOR
JPN
NOR
UKR
MNG
LKA
LTU
NZL
UKR
NLD
UKR
CAN
FIN
CHN
RUS
NOR
AUT
JPN
FRA
PHL
FRA
MAR
LAO
HRV
MNG
EST
CHE
MNG
ROM
UGA
MNG
GRC
SYR
MAR
ARG
HUN
CHE
CHN
CHN
GHA
NLD
LAO
HUN
PER
LAO
HRV
BRA
UGA
HUN
CZE
JPN
MEX
BGR
EST
KAZ
EST
SLV
BRA
RUS
TZA
BEL
UKR
NZL
USA
CHN
ARE
PHL
ROM
LKA
FRA
LVA
LTU
BRA
MAR
MOZ
TUR
JPN
LTU
BRA
LVA
CHN
GEO
CHE
CHE
SWE
CHN
ARG
JOR
LAO
NLD
CAN
USA
MNG
BGR
LVA
FIN
UKR
TZA
BEL
MEX
IND
SWE
JPN
GRC
PER
MYS
LAO
BRA
HUN
ARE
SYR
IND
HRV
MAR
HRV
NZL
LVA
DEU
CHE
DEU
PER
VNM
DEU
NLD
PER
BRA
UKR
ARG
BEL
PHL
GRC
CHN
CHN
HUN
DNK
IND
GEO
VNM
EST
RUS
MEX
CHN
KAZ
SLV
MAR
LAO
GHA
PER
BRA
BRA
GEO
LKA
PHL
CHN
RUS
NLD
JPN
JOR
PER
LAO
MNG
UKR
ESP
MAR
LTU
CHN
UKR
UKR
MEX
NLD
ROM
PAN
RUS
LVA
LAO
PAN
ARG
AUT
MKD
IND
POL
DNK
ARG
BEN
ARE
LVA
BEL
PAN
RUS
JOR
TUR
DNK
JOR
GHA
SYR
LVA
CHN
SLV
MEX
USA
CHN
RUS
PER
MEX
MEX
ESP
GRC
SLV
CHE
RUS
DEU
SYR
MNG
LTU
MDA
ROM
SLV
KAZ
ARG
RUS
MEX
EST
MEX
SYR
MEX
IND
IND
MNG
LVA
UKR
BEN
PER
SWE
PER
ARG
BRA
UKR
GHA
MEX
TUR
MAR
ESP
UKR
CHN
KAZ
DNK
CHN
ARG
TZA
BEN
SLV
JOR
FRA
CZE
BRA
LVA
NPL
AUT
TUR
SLV
MYS
USA
USA
TZA
MKD
GRC
BEN
UKR
DNK
BEN
TZA
GEO
USA
CMR
CHN
MDA
POL
LVA
GEO
BEN
MOZ
USA
USA
JOR
ARG
BEL
GRC
THA
GHA
LKA
MYS
PAK
VNM
TZA
MOZ
DNK
ESP
ITA
TZA
MYS
FRA
CMR
BRA
MAR
GEO
MEX
IND
MYS
PHL
POL
MOZ
PHL
ARG
RUS
IND
MOZ
MEX
MEX
THA
LVA
TUR
ITA
DEU
THA
DNK
NPL
UGA
FRA
TZA MYS
RUS
LAO
PAN PHL
PHL
DNK
IND
PHL HRV
SRB
SEN PHL
ITA
CMR
ITA
ESP ESP PHL
SEN
ITA
CMRITA
-2
0
2
Years of Education
IND
4
coef = .12334688, (robust) se = .02273828, t = 5.42
PHL
-2
CMRTHA
DEU
JOR
SEN
THA
DEU
PAN
SLV
MYS
PER
CMR
LAO
GEO
MYS
MEX
PHL
MOZ
MEX
LKA
CMR
MEX
PAN
MOZ
CMR
MAR
MEX
SLV
IND
RUS
POL
ARG
PER
ARG
PHL
POL
BRA
MYS
GHA
MEX
IND
RUS
ARE
UKR
BEL
NPL
MYS
MEX
ARE
PAN
SYR
RUS
CHE
PHL
MEX
POL
SLV
PER
IND
ESP
BEN
RUS
RUS
BEN
PER
UKR
IND
UKR
MEX
BRA
PER
MEX
IND
GEO
IND
RUS
CAN
ARG
MYS
RUS
ARG
RUS
JPN
PER
PHL
BRA
DNK
RUS
FRA
BRA
DNK
RUS
USA
HRV
HRV
CHN
CHN
ITA
DNK
USA
DNK
TUR
PER
PHL
DNK
CMR
BRA
SYR
SYR
LAO
BRA
HRV
USA
BEN
HRV
FRA
RUS
AUT
CHN
RUS
POL
MEX
SYR
RUS
HRV
GEO
MAR
NPL
POL
RUS
RUS
JOR
POL
MKD
FRA
JPN
BGR
MAR
VNM
PER
TZA
BRA
BEN
HRV
GRC
HRV
HRV
DNK
GEO
CAN
UKR
BEN
DNK
MAR
ITA
DNK
RUS
NOR
BRA
ITA
BEL
CHN
CHN
GHA
CHE
PER
RUS
MEX
EST
GRC
BEN
POL
ESP
FRA
SYR
NOR
UKR
GHA
SYR
MYS
SWE
FRA
MNG
RUS
MNG
USA
EST
USA
NLD
LVA
HRV
HUN
JOR
CHN
USA
PHL
RUS
ESP
LVA
RUS
PAN
TUR
CZE
CHE
USA
ROM
FRA
CHE
USA
POL
MEX
TZA
JPN
NOR
ARG
CHE
BEL
GRC
JPN
HUN
VNM
JPN
SYR
UGA
ESP
IND
LTU
VNM
GRC
SEN
LAO
IND
MNG
TZA
IND
NLD
ESP
LAO
POL
UGA
SRB
MYS
BRA
CHN
KAZ
ROM
CZE
LKA
TUR
USA
AUT
MOZ
CHN
MNG
ROM
RUS
EST
GEO
USA
MAR
UKR
CZE
TUR
CHE
NZL
PER
USA
USA
FIN
JPN
FRA
CHN
LTU
JPN
TZA
ARE
GHA
EST
NZL
BRA
CHN
PER
GHA
ARG
JOR
MEX
MKD
GRC
LVA
POL
NZL
UKR
KAZ
USA
NLD
MEX
USA
LVA
HUN
AUT
POL
PAK
UKR
NZL
SWE
MOZ
JPN
EST
JPN
ROM
CZE
NOR
TZA
UKR
CHE
NOR
NLD
USA
KAZ
CAN
NZL
JPN
MNG
ITA
CHE
CMR
SLV
KAZ
SWE
JOR
GEO
IND
LTU
RUS
NLD
ITA
CAN
SLV
BEL
NLD
PAK
NOR
ESP
NZL
ESP
IRL
FRA
LAO
ESP
NLD
LVA
UKR
NOR
NZL
MAR
NOR
ESP
EST
LVA
BGR
NZL
CAN
RUS
ESP
ARG
NOR
LAO
LVA
GRC
JOR
TZA
ESP
USA
MNG
LVA
NOR
LTU
ITA
SLV
MEX
JPN
FIN
LTU
ARG
FIN
LVA
BIH
IND
ITA
RUS
BGR
CHN
POL
AUT
CAN
USA
MNG
TUR
LAO
PAK
RUS
GRC
ESP
ITA
USA
MEX
TZA
MKD
PAN
ARG
SWE
JPN
TUR
LVA
TZA
EST
LVA
HUN
NZL
LVA
LVA
NLD
MNG
TZA
PER
TZA
LAO
BEL
BEL
LVA
LVA
BRA
IRL
CHN
NZL
MAR
LAO
LAO
SWE
LVA
BRA
BGR
TZA
UKR
PAK
ROM
PHL
UKR
EST
JOR
BRA
VNM
FRA
RUS
IND
CAN
CHE
BGR
ESP
GEO
JPN
NLD
LTU
EST
TZA
AUT
NOR
LVA
EST
VNM
RUS
JPN
NPL
LVA
HUN
AUT
FRA
SYR
UKR
ARG
AUT
MKD
LVA
LVA
SYR
LVA
PER
LVA
CHN
MKD
BRA
LAO
JPN
SLV
JPN
HUN
GRC
MEX
LTU
RUS
JPN
TUR
ITA
JPN
FIN
EST
TZA
BEL
HRV
MAR
LTU
USA
MNG
UGA
CHE
EST
UGA
NZL
BGR
SWE
LVA
VNM
GRC
UKR
JPN
CHN
ARG
JPN
PER
ARG
IND
USA
IND
GHA
MNG
VNM
TUR
SRB
IND
LAO
CAN
FRA
IND
PHL
BEL
ROM
TZA
TUR
VNM
IND
TUR
BRA
LTU
HRV
CHN
SYR
CAN
JPN
RUS
ITA
EST
TZA
MNG
GRC
BEL
JPN
MNG
RUS
MKD
CHE
PHL
FRA
SYR
USA
ARE
LAO
MAR
USA
NLD
ROM
TUR
CAN
USA
BRA
IND
GRC
KAZ
GHA
ITA
BRA
MNG
PHL
BRA
LAO
CHN
CZE
IND
PHL
HUN
PER
MAR
CHN
DNK
GEO
AUT
UKR
DNK
PHL
JPN
HRV
IND
LKA
NOR
UKR
MAR
NOR
LKA
HRV
BRA
BRA
CHE
LAO
DNK
MNG
RUS
IND
CZE
IND
SYR
BEL
MNG
CMR
ARG
MYS
PAN
MYS
BEN
GRC
NOR
ARE
LKA
CHN
ARG
RUS
JPN
MOZ
UKR
RUS
GHA
BEN
USA
PAN
BRA
RUS
PER
CAN
LAO
CHN
GHA
PER
LAO
IND
MEX
IND
UKR
BEL
IND
SLV
ITA
LKA
ESP
MAR
TZA
SLV
NPL
SLV
CHE
PHL
USA
ITA
THA
SLV
JOR
CHN
RUS
BRA
BRA
HRV
HRV
UKR
JOR
HRV
UKR
HRV
ARG
ESP
GEO
ARG
SLV
RUS
MEX
RUS
GHA
IND
ARG
MEX
UKR
MOZ
MEX
ESP
MOZ
FRA
JOR
POL
JOR
MYS
ARE
SLV
NPL
FRA
PHL
SYR
PHL
PER
PER
MEX
ARG
JOR
MEX
PER
GEO
MAR
USA
BRA
DNK
MOZ
DNK
MEX
DNK
PER
PER
SYR
RUS
HRV
FRA
MEX
PAN
MEX
CHN
PER
MEX
BEN
RUS
MYS
RUS
CMR
GEO
CMR
RUS
MEX
MYS
RUS
THA
MYS
DEU
RUS BEN
RUS
POL
THA
SEN
POL
PAN MEX
DEU DEU
CMR
DEU
DEU
DEU
-4
coef = .29672036, (robust) se = .03135326, t = 9.46
-4
DEU
-5
-4
SEN
ITA
ITA
DEU
Employees/Firm
0
5
10
4
Years of Education and Participation in the Official Economy
Figure 7
PERLAO
LAO
PER
PER
LAO PER
VNM
PER
ITA VNM
GHA
LVA
PER
SLV
LVA
ITA
PER
SLV
EST
ITA
PER
LAO
MEX
PER
TZA
ITA
VNM
LVA
TZA
PER UGA
SLV
LVA
PER MEX
NOR
TZA
GHA
ITA
PER
MEX
MNG
MEX
PAK
ARG
EST
MEX
JPN
TZA
MEX
MEX
EST
GHA
NZL
GEO
TZA
LVA
ITA
GEO
ARG
PAK
MOZMNG
VNM
MOZ
ARG
NPL
MEX
TZA
MEX
LVA
PHL
GEO
ARG
GHA
JPN
LAO
JPN
NOR
LTU
ARG
MNG
MNG
MEX
NOR
ARG
TZA
LVA
MKD
PHL
JPN
SLV
GEO
SLV
LVA
LTU
EST
THA
NZL
ARG
NZL
GEO
NZL
GHA
PHL
MNG
LVA
LTU
NOR
LKA
NZL
KAZ
SLV
NOR
JPN
PER
KAZ
MNG
CAN
TZA
GHA
LAO
NZL
PHL
NOR
CAN
MEX
MNG
MEX
JPN
CAN
JPN
PHL
EST
NPL
LVA
THA
ARG
ARG
LVA
MOZ
JPN
UKR
LVA
UKR
JPN
UKR
JPN
LVA
CAN
CAN
CAN
JPN
MEX
JPN
MNG
NZL
JPN
VNM
CAN
MKD
PHL
MEX
MEX
TZA
MNG
GEO
MOZ
MEX
NOR
USA
NZL
MNG
KAZ
EST
UKR
CAN
JPN
USA
JPN
UKR
MEX
LVA
USA
UGA
CHN
THA
TZA
USA
CHN
JPN
MKD
JPN
USA
MOZ
LKA
USA
USA
UKR
CHN
CHN
NPL
USA
JPN
LTU
EST
CHN
CHN
CHN
UKR
PHL
EST
ITA
USA
LKA
JPN
CHN
ARG
CHN
ARG
CAN
MKD
UGA
THA
SLV
CHN
USA
GEO
USA
USA
USA
CHN
USA
NOR
ARG
USA
USA
UKR
CHN
MKD
UKR
UKR
LAO
ARG
UKR
MOZ
LKA
GHA
USA
USA
PHL
CHN
PHL
CHN
CHN
CHN
JPN
JPN
LVA
CHN
KAZ
LTU
MNG
UKR
JPN
USA
ARG
JPN
USA
LTU
NOR
USA
PHL
USA
MNG
NOR
CHN
UKR
UKR
USA
NZL
CHN
MNG
CHN
USA
CHN
ARG
PHL
UKR
USA
CHN
UKR
USA
JPN
NOR
SLV
NOR
JPN
JPN
ITA
TZA
UKR
PHL
JPN
JPN
USA
JPN
MEX
UKR
LKA
USA
MEX
NOR
USA
MOZ
LVA
JPN
LKA
JPN
GEO
MNG
USA
CHN
UKR CHN CHN
JPN
JPN
UKR
GEO
JPN
JPN
NOR
NPL
KAZ
TZA
MKD
JPN
NOR
JPN
ARG
MNG
PHL
UKR
MKD
MNG
PHL
LTU
LKA
LTU
NOR
ARG
USA
PHL
JPN
ITA
UKR
MEX
JPN
PER
USA
ARG
LVA
JPN
MEX
TZA
TZA
JPN
ARG
LTU
PER
JPN
MNG
LAO
MKD
JPN
ARG
MEX
NZL
MEX
LVA
TZA
NOR
PHL
JPN
SLV
LTU
MNG
LVA
SLV
EST
PAK
TZA
MEX
TZA
PER
NZL
MEX
SLV
MEX
MEX
LAO
GEO
NOR
LAO
KAZ
LVA
ARG
MOZ
NPL
PER
SLV
TZA
EST
LVA
NZL
ARG
THA
SLV
PHL
MOZ
ITA
PAK
MNG
UGA
ARG
MNG
NOR
EST
ITA
NZL
MEX
GHA
GHA
LVA
EST
PER MEX SLV
TZA
PER
MEX
PER
LVA
GEO
GHA PER ITA
VNM
ITA CAN LAO
TZA
PER LAO LVA
VNM
LAO
PER
PER
-2
-1
0
1
Years of Education
2
coef = .24454543, (robust) se = .03739797, t = 6.54
3
44
Conclusion
45
Conclusion

Education is the only variable that explains a substantial amount of regional variation in
income and labor productivity.
 Education influences regional development through education of workers, education of
entrepreneurs, and substantial regional externalities.
 Better educated regions have larger, more productive firms, and higher labor force
participation.

The evidence is hard to reconcile with the standard Cobb-Douglas production function.
 Our Lucas-Lucas model combines allocation of talent, externalities, and migration.

Central Message:
1. Private returns to worker education are modest; but
2. Private returns to entrepreneurial education and social returns to education are large.

Our data suggest that education increases the supply of entrepreneurs as well as improves
46
the scope for the exchange of ideas ↔ economic development.
Random Coefficients Estimation
RANDOM COEFICIENTS (COUNTRY LEVEL); RANDOM EFFECTS (REGION); DEMEANED (COUNTRY LEVEL)
Logarithm of Sales per employee
(2)
(3)
(1)
(4)
(5)
Logarithm of Wages per employee
(6)
(7)
(8)
Years of Schooling in the Region
0.0775
(0.0198)
a
0.0708
(0.0188)
a
0.0809
(0.0336)
b
0.0813
(0.0337)
b
0.0775
(0.0198)
a
0.0708
(0.0188)
a
0.0809
(0.0336)
b
0.0813
(0.0337)
Ln(Population in the Region)
0.0514
(0.0416)
0.0447
(0.0394)
0.1014
(0.0501)
b
0.1028
(0.0504)
b
0.0514
(0.0416)
0.0447
(0.0394)
0.1014
(0.0501)
b
0.1028
(0.0504)
Years of Schooling of managers
0.0609
(0.0078)
a
0.0412
(0.0072)
a
0.0586
(0.0102)
a
0.0599
(0.0103)
a
0.0609
(0.0078)
a
0.0412
(0.0072)
a
0.0586
(0.0102)
a
0.0599
(0.0103)
.
.
0.1467
(0.0081)
a
.
.
‐0.0091
(0.0056)
.
.
0.1467
(0.0081)
a
.
.
‐0.0091
(0.0056)
a
0.0259
(0.0103)
0.0263
(0.0104)
0.0352
(0.0072)
a
0.0312
(0.0073)
a
0.0259
(0.0103)
0.0263
(0.0104)
a
.
.
0.3492
(0.0065)
a
0.3492
(0.0064)
a
.
.
.
.
a
.
.
.
.
0.3073
(0.0048)
a
0.3079
(0.0048)
Ln(Employees)
Years of Schooling of workers
0.0352
(0.0072)
a
0.0312
(0.0073)
Ln(Expenditure on energy/employee)
0.3492
(0.0065)
a
0.3492
(0.0064)
.
.
.
.
.
.
0.3073
(0.0048)
a
0.3079
(0.0048)
Ln(Property, Plant, Equipment / employees)
b
b
b
b
b
a
b
a
Constant
5.3025
(0.7363)
a
5.2351
(0.7033)
a
4.2483
(0.9891)
a
4.2297
(0.9930)
a
5.3025
(0.7363)
a
5.2351
(0.7033)
a
4.2483
(0.9891)
a
4.2297
(0.9930)
Observations
Log Likelihood
Chi Squared
Prob > Chi Squared
13,248
‐18,655
4,325
0
13,248
‐18,493
4,788
0
19,305
‐26,361
6,601
0
19,305
‐26,360
6,603
0
13,248
‐18,655
4,325
0
13,248
‐18,493
4,788
0
19,305
‐26,361
6,601
0
19,305
‐26,360
6,603
0
Note: a = significant at the 1% level, b = significant at the 5% level, and c = significant at the 10% level. a
47